用于病理性近视识别的高效金字塔通道注意力网络
Efficient Pyramid Channel Attention Network for Pathological Myopia Recognition
AI总结:
针对病理性近视识别中现有网络未利用病灶分布先验的问题,提出高效金字塔通道注意力模块并构建EPCA-Net,结合预训练-微调范式与适配器,在自建PM-fundus基准上取得优于现有方法的性能。
摘要:
病理性近视(Pathological myopia, PM)是全球范围内导致视力受损的首要眼部疾病。临床上,PM的病灶在眼底图像上呈现全局-局部分布的特征,这一特征对辅助临床医生诊断PM具有重要作用。然而,现有大多数深度神经网络专注于设计复杂架构,却很少挖掘PM的病灶分布先验知识。为解决这一问题,我们提出了高效金字塔通道注意力(Efficient Pyramid Channel Attention, EPCA)模块,该模块通过金字塔池化和多尺度上下文融合,充分挖掘PM临床病灶先验的潜力。随后,我们通过堆叠一系列EPCA模块,构建了用于基于眼底图像自动识别PM的EPCA-Net。此外,受近期预训练-微调范式的启发,我们尝试将预训练的自然图像模型适配到PM识别任务中:冻结预训练模型,将EPCA及其他注意力模块作为适配器。另外,我们通过收集公开数据集中的PM眼底图像,构建了一个名为PM-fundus的PM识别基准。综合实验证明,在PM识别任务中,我们的EPCA-Net性能优于现有最先进方法。实验结果还表明,与部分采用传统微调范式的已有方法相比,我们基于预训练-微调范式的方法以更少的可训练参数取得了具有竞争力的性能,这为在医疗数据有限的场景下,利用更多自然图像基础模型解决PM识别任务提供了可能。
英文摘要:
Pathological myopia (PM) is the leading ocular disease for impaired vision worldwide. Clinically, the characteristic of pathology distribution in PM is global-local on the fundus image, which plays a significant role in assisting clinicians in diagnosing PM. However, most existing deep neural networks focused on designing complex architectures but rarely explored the pathology distribution prior of PM. To tackle this issue, we propose an efficient pyramid channel attention (EPCA) module, which fully leverages the potential of the clinical pathology prior of PM with pyramid pooling and multi-scale context fusion. Then, we construct EPCA-Net for automatic PM recognition based on fundus images by stacking a sequence of EPCA modules. Moreover, motivated by the recent pretraining-and-finetuning paradigm, we attempt to adapt pre-trained natural image models for PM recognition by freezing them and treating the EPCA and other attention modules as adapters. In addition, we construct a PM recognition benchmark termed PM-fundus by collecting fundus images of PM from publicly available datasets. The comprehensive experiments demonstrate the superiority of our EPCA-Net over state-of-the-art methods in the PM recognition task. The results also show that our method based on the pretraining-and-finetuning paradigm achieves competitive performance through comparisons to part of previous methods based on traditional fine-tuning paradigm with fewer tunable parameters, which has the potential to leverage more natural image foundation models to address the PM recognition task in limited medical data regime.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。